Food scientists at the National University of Singapore (NUS) have developed a machine learning model that maps 991 aromatic plants and identifies scent compounds potentially beneficial for sleep. This groundbreaking study was published on July 6, 2026, and aims to provide alternatives to conventional sleep medications.
Machine Learning Revolutionizes Sleep Aid Discovery
Led by Assistant Professor Zhang Dachuan, the NUS team analyzed over 2,300 scent molecules to identify those that promote sleep. Their research, featured on the cover of the journal Digital Discovery, demonstrates the effectiveness of a data-driven approach in the search for natural sleep aids.
The team compiled a curated library of 2,391 scent molecules derived from 991 aromatic plants. By employing a machine-learning model, they achieved an impressive 96.1% accuracy in distinguishing compounds that promote sleep from inactive ones. This model highlighted numerous high-potential candidates for further investigation.
Aromatic Plants with Sleep-Promoting Properties
Among the findings, four commercially available molecules—carvacrol, safranal, vanillin, and methyl eugenol—were tested for their effects on sleep. The results showed that these compounds reduced wakefulness and increased non-rapid eye movement sleep, a crucial restorative sleep stage.
The study revealed that these molecules interact with GABA receptors, which are significant in the brain's calming signaling system and a common target for conventional sleep medications. This insight lays the groundwork for understanding how specific plant aromas might influence sleep pathways.
Future Directions in Sleep Research
Beyond individual molecules, the research team prioritized plant families rich in promising scent compounds. Families such as Asteraceae, Lamiaceae, and Lauraceae were highlighted, with species like lavender and perilla noted for further study. The findings serve as a practical map for future research rather than a direct solution for sleep issues.
Assistant Professor Zhang emphasized the need for stronger evidence to connect calming plant aromas with sleep improvement. “Our aim is to transform traditional knowledge and scattered chemical data into a practical map that can guide the development of safer and more targeted sleep-related products in the future,” he stated.
Future investigations will focus on long-term safety, the interactions between scent molecule mixtures, and the biological effects observed in broader studies. This research highlights the potential of AI in accelerating the discovery of natural ingredients for health products, functional foods, and wellness applications.
🤖 This article was rewritten by Feed and Figures' editorial AI from a report originally published by Phys.org. Facts and quotes are preserved from the original; the rewrite focuses on clarity and structure. For the unedited original, see the source link below.
